A batch can pass every specification and still tell you that something is changing. Trending is how a quality system hears it.
Why the guide asks for trending
Chapter 6 of the EU GMP Guide says that data such as test results, yields and environmental controls should be recorded in a way that permits trend evaluation, and that results of parameters identified as quality attributes or as critical should be trended. It also calls for any out-of-trend or out-of-specification data to be addressed and investigated. Chapter 1 adds that the pharmaceutical quality system should include monitoring of process performance and product quality, and the product quality review is expected to evaluate trends across the year.
OOT is not the same as OOS
An out-of-specification result breaches an approved limit. An out-of-trend result is a value that is within specification but atypical compared with historical data, for example a drift in assay across consecutive batches or a stability result falling faster than expected. Regulations do not give a single numerical definition of OOT, so each company must define and justify its own criteria.
Building a practical trending programme
- Select what to trend. Use risk assessment to choose critical quality attributes, critical process parameters, in-process controls, yields, environmental monitoring and stability data.
- Define statistical rules. Control charts with agreed rules, such as points beyond control limits or a run of consecutive points on one side of the mean, give objective alert triggers. Ensure the statistical method suits the data distribution.
- Set alert and action levels. Distinguish between a signal that requires review and one that requires a formal investigation.
- Assign ownership and frequency. Name who reviews, how often and where the outcome is recorded.
Stability and ongoing studies
Stability data are a frequent source of OOT signals. The ongoing stability programme in Chapter 6 exists to detect issues in the marketed product, and the summary of data should be reviewed periodically. Regression or other trend analysis of time-point data can reveal a shortened shelf life before a limit is crossed.
Avoiding common weaknesses
- Trending performed only at the annual review, when it is too late to act.
- Charts produced but never reviewed, signed or acted upon.
- Limits set from too few batches, or never recalculated after process changes.
- No link between OOT signals and the deviation or change control systems.
A short self-check
- Is there a written definition of an out-of-trend result, with a justification for the criteria?
- Are control limits recalculated after a process change?
- Can you show an example where a trend signal led to action before any specification failure?
How PQRA helps
PQRA supports sites in designing trending programmes, choosing attributes and statistical rules, drafting OOT procedures and reviewing how trend signals flow into deviations, CAPA and the product quality review. We work with quality assurance, QC and production teams so that the programme is proportionate and usable.
Contact PQRA to discuss how we can support your quality system.


No responses yet